Blog
Back to Blog
March 16, 2026

Melt Curve Analysis in qPCR: What It Tells You and When to Worry

A melt curve with a single, sharp peak at the expected temperature means your primers amplified one product. That's the whole point of melt curve analysis — it's your post-amplification check that the fluorescence you measured during cycling actually came from your target and not from primer dimers, off-target amplicons, or genomic DNA contamination. If you're running SYBR Green (or any intercalating dye) without melt curves, you're essentially trusting your Ct values on faith.

The practical rule: one primer pair should give one peak, at a consistent Tm across all your samples, with nothing in the NTC. When that's what you see, move on to your analysis. When it's not — a shoulder, a second peak, a shifted Tm, or signal in the no-template control — that's when you need to dig in. The rest of this post is about what those deviations mean and what to do about them.

How Melt Curves Work

After amplification finishes, the instrument ramps the block temperature from ~60°C to ~95°C in small increments (typically 0.5°C steps on a CFX96 or QuantStudio, 0.2°C on a LightCycler 480 in high-resolution mode). At each step, it measures fluorescence. As double-stranded DNA denatures, the intercalating dye (SYBR Green I, EvaGreen, or equivalent) is released and fluorescence drops. The raw data is a sigmoidal decay curve — fluorescence vs. temperature.

What you actually look at is the negative first derivative of that curve, −dF/dT. This converts the gradual fluorescence drop into a peak at the temperature where the rate of denaturation is highest. That peak is the melting temperature (Tm) of the product. A 100-200 bp amplicon from a well-designed primer pair typically gives a Tm between 78°C and 88°C, depending on GC content and length. The peak should be narrow (spanning ~2-3°C at half-height) and symmetrical.

If you've never looked at the raw fluorescence curve, it's worth doing at least once — it helps you understand why the derivative plot looks the way it does and why broad or asymmetric peaks indicate heterogeneous products.

What a Good Melt Curve Looks Like

For a standard qPCR assay — say GAPDH with a 120 bp amplicon run with PowerUp SYBR Green — you want to see:

If all of that checks out, your data is clean. Spend your time on the analysis, not the QC.

Primer Dimers: The Most Common Problem

Primer dimers show up as a second peak at a lower Tm than your target — usually 72-78°C for a dimer, vs. 80-88°C for a real amplicon. They're short (20-40 bp of concatenated primer sequence), so they melt at a lower temperature.

Where you'll see them:

  1. NTC wells. This is expected at low levels. If your NTC shows a dimer peak with an apparent Ct of 36-40, that's generally fine as long as your experimental samples are well below that (Ct <30 or so). The dimer forms because primers have 40 cycles with nothing else to do.
  2. Low-template samples. If you're quantifying rare transcripts and your Ct values are 32-35, primer dimers can start competing with real amplification. The melt curve is your only way to know whether that late Ct is real signal or artifact.
  3. All wells, including high-template ones. This means your primers have significant complementarity and are forming dimers even when target is abundant. Redesign the primers or, at minimum, increase the annealing temperature by 2-3°C to see if you can disfavor the dimer.

The critical question is whether the dimer contributes to the Ct value you're reporting. If the melt curve shows 95% of the fluorescence in the target peak and 5% in a dimer peak, your Ct is probably fine — the dimer formed late in cycling after the Ct was already determined. If the two peaks are comparable in height, that Ct is compromised.

Shoulders, Double Peaks, and Shifted Tm Values

A shoulder on the main peak — where one side of the peak is broader than the other or has a slight inflection — usually means one of two things: a splice variant or pseudogene amplicon of similar size, or a heteroduplex product. Heteroduplexes form when amplicons from slightly different sequences (e.g., two closely related family members) reanneal as mismatched duplexes during the melt ramp. These melt at slightly lower Tm than perfect homoduplexes, creating the shoulder.

What to do: Run the product on a 2-3% agarose gel or, better yet, on a Bioanalyzer/TapeStation. If you see one clean band, the shoulder is likely heteroduplex and the assay is probably fine for relative quantification. If you see two bands, your primers hit two targets and you need more specific primers or a TaqMan probe to resolve them.

Two clearly distinct peaks mean two different amplicons. Check your primer specificity in silico (Primer-BLAST against the RefSeq transcriptome) and run a gel. This assay isn't giving you reliable data for either target.

Tm shift between samples — for example, your treated samples melt at 83°C and your controls melt at 84.5°C — can indicate:

NTC Signal: When to Worry

A clean NTC (no amplification, no melt peak) is ideal but not always what you get. Here's a practical decision tree:

Melt Curves and TaqMan: Why You Don't Need Them

If you're running TaqMan (hydrolysis probe) assays, you don't run melt curves. The probe provides the specificity — it only generates signal if it hybridizes to the correct internal sequence and gets cleaved. Primer dimers can still form, but they don't generate fluorescence because there's no probe to cleave. This is one of the genuine advantages of probe-based chemistry for targets where SYBR specificity is hard to achieve (highly homologous gene families, processed pseudogenes, low-abundance targets).

That said, SYBR is cheaper, faster to set up, and perfectly fine for the vast majority of gene expression work where primers are specific. Just run the melt curve.

High-Resolution Melting: A Different Application

High-resolution melt (HRM) analysis uses the same principle but with saturating dyes (like EvaGreen or SYTO 9), finer temperature increments (0.1-0.2°C), and more sensitive optics. HRM can distinguish amplicons that differ by a single nucleotide — making it useful for SNP genotyping, mutation scanning, and methylation analysis. The LightCycler 480 and Rotor-Gene Q both have dedicated HRM modes. This is a different application from routine melt curve QC, and the analysis is more involved. Don't confuse the two: standard melt curves tell you if you have one product; HRM tells you if that product varies between samples at the sequence level.

Quick Checklist

Before you trust a SYBR Green qPCR dataset, confirm these from the melt curves:

  1. Single peak per primer pair, consistent Tm (±0.5°C) across all samples.
  2. NTC wells show no peak at the target Tm.
  3. No secondary peaks or shoulders in experimental wells.
  4. Replicate Ct values agree within 0.5 Ct (this is amplification data QC, not melt curve per se, but check both together).

If any of those fail, flag the affected wells before calculating ΔCt or fold change. Reporting fold changes from wells with ambiguous melt curves is how irreproducible results get into papers.

If you're processing multiple plates or large experiments, checking every melt curve manually gets tedious fast. VoilaPCR flags wells with abnormal melt profiles automatically when you upload your data, so you can focus on the biology instead of scrolling through 384 derivative plots.